speckit-agent-context-update
Pass
Audited by Gen Agent Trust Hub on Oct 23, 2026
Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill executes local bash and PowerShell scripts located at
.specify/extensions/agent-context/scripts/to update the agent's context file. - [INDIRECT_PROMPT_INJECTION]: The skill manages a section of the agent's persistent instruction files, which could be exploited if malicious content is placed in the project specification files being referenced. * Ingestion points: Reads configuration from
.specify/extensions/agent-context/agent-context-config.ymland content fromspecs/*/plan.md. * Boundary markers: Uses<!-- SPECKIT START -->and<!-- SPECKIT END -->to delimit the injected section. * Capability inventory: Executes local scripts and modifies persistent instruction files (e.g.,CLAUDE.md,.github/copilot-instructions.md). * Sanitization: No sanitization of the plan file content or path is mentioned before injection into the instruction file.
Audit Metadata